Note: Descriptions are shown in the official language in which they were submitted.
METHOD FOR OPERATING A WIND TURBINE
The invention relates to a method for operating a wind turbine, and to a wind
turbine
designed for performing this method and to a corresponding computer program
product.
Wind turbines are known from the prior art. They generally comprise a rotor,
which is
arranged rotatably on a nacelle, the nacelle in turn being arranged rotatably
on a tower. The
rotor drives a generator, possibly via a rotor shaft and a gear mechanism. A
wind-induced
io rotational movement of the rotor can thus be converted into electrical
energy, which can then
be fed via converters and/or transformers ¨ depending on the type of
construction of the
generator also at least partially directly ¨ into an electric grid. The rotor
comprises a number
(generally three) of rotor blades, which extend substantially radially from
the rotor axis and
are fastened pivotably with respect to a rotor hub in order to set the angle
of attack of the
rotor blades.
The rotor blades of the wind turbines are often adjustable with regard to
their blade angle
(pitch adjustment), whereby the angle of attack of the individual rotor blades
can also be
changed during operation. In a partial load range between the cut-in speed,
from which the
rotor can begin to turn, and the rated wind speed, from which the wind turbine
feeds its rated
power into the electric grid, the blade angle is chosen so as on the one hand
to obtain the
maximum possible wind yield and on the other hand to be certain of avoiding a
stall at one or
more rotor blades. In the full load range, with a wind speed beyond the rated
wind speed, the
blade angles are set such that the rotor only rotates at the prescribed
maximum speed.
In the partial load range, the blade angles are often controlled on the basis
of a characteristic
curve, in dependence on the dimensionless tip speed ratio A, as it is known,
which is
obtained from the quotient of the speed of the rotor blade tips and the wind
speed. The
characteristic curve is chosen such that, when it is maintained in the partial
load range, the
optimum power coefficient of the rotor or the rotor blades is achieved, which
in turn means
that there is an optimum wind yield. In the full load range, on the other
hand, a blade angle
deviating from this characteristic curve is set, whereby the power coefficient
of the rotor is
reduced.
1
CA 3009204 2018-06-22
Depending on the location where a wind turbine is installed, there is the risk
of ice forming on
the rotor blades, which changes the properties of the flow passing around the
rotor blades. In
particular, there is an increase in the risk of stalls ¨ if the control
remains the same. Such a
stall means that there is a sudden drop in the energy generation and, as a
consequence, a
considerable mechanical loading for the wind turbine. Also, such a sudden drop
in the energy
that is fed in also has an adverse effect on the stability of the electric
grid. A stall must
therefore be avoided as far as possible.
US Patent US 8,096,761 B1 has discovered that, in cases of poorer aerodynamics
of the
rotor blades ¨ for example due to the formation of ice ¨ a characteristic
curve used for
io controlling the blade angles in the partial load range can be used to
ascertain a comparable
(if not identical) curve that indicates for each of the different operating
states a minimum
blade angle below which there is an increased risk of a stall occurring. As a
consequence,
the wind turbine control is designed such that, when there are poorer
aerodynamics of the
rotor blades, the angle of the blades does not go below the minimum blade
angle for the
respective operating state.
A disadvantage of this prior art is that, although a stall at the rotor blades
when there are
poorer aerodynamics can be avoided, it may be that the wind turbine cannot
deliver the
maximum possible yield for the turbine because of the minimum blade angle
fixed as a lower
limit. In particular in the case of turbines installed at locations where the
formation of ice on
zo the rotor blades can be expected over a longer period of time, the loss
of yield due to
maintaining a minimum blade angle, which is only concerned with the certain
avoidance of
stalls, may be considerable.
The object of the present invention is to provide a method for operating a
wind turbine, and
also a wind turbine and a computer program product, with which the
disadvantages of the
prior art no longer occur, or at least only to a reduced extent.
This object is achieved by a method according to the main claim, and also a
wind turbine and
a computer program product according to the alternative independent claims.
Advantageous
developments are the subject of the dependent claims.
Accordingly, the invention relates to a method for operating a wind turbine
comprising a rotor
with a number of rotor blades that can be set with regard to the blade angle
and a detection
of the formation of ice on the rotor blades, the blade angle setting in normal
operation being
2
CA 3009204 2018-06-22
performed on the basis of a standard characteristic curve, in dependence on a
characteristic
number that can be ascertained during the operation of the wind turbine, and
in the case
where the formation of ice is detected taking place according to the following
steps:
f) operating the wind turbine on the basis of an initial special
characteristic curve for
the blade angles of a rotor blade or all the rotor blades, in dependence on a
characteristic number that can be ascertained during the operation of the wind
turbine;
g) recording a first power curve for a prescribed time period;
h) changing the special characteristic curve;
i) recording a further power curve for a prescribed time period; and
j) checking whether the last ascertained further power curve
represents an
optimum:
- if so, operating the wind turbine on the basis of the optimum special
characteristic curve (that is to say, depending on the definition of the
special characteristic curve in step (a), either for the one rotor blade or
all
the rotor blades) on which the last ascertained further power curve is
based;
- if not, iteration from step c).
The invention also relates to a wind turbine comprising a rotor with a number
of rotor blades
zo .. that can be set with regard to the blade angle, which is arranged
rotatably on a nacelle
arranged rotatably on a tower, and with a generator arranged in the nacelle
for the
conversion of wind energy acting on the rotor into electrical energy, and a
control device for
controlling the wind turbine and its components, the control device being
designed for
carrying out the method according to the invention.
3
CA 3009204 2018-06-22
The invention also relates to a computer program product comprising program
parts which,
when loaded in a computer, are designed for carrying out the method according
to the
invention.
First of all, some of the terms used in connection with the invention are
explained.
The term "power curve" refers to the quantitative relationship between the
wind speed and
the electrical power generated.
For "recording a power curve", values registered for wind speed and generated
electrical
power are brought together in pairs to form data points. If the recording is
performed over a
sufficiently long time period with changing wind speeds, a power curve is
obtained from the
io individual data points. However, it is also possible that only a small
number of data points are
registered, from which a power curve can then be extrapolated if need be. In
an extreme
case ¨ in particular when the wind stays the same over a longer time period ¨
it is also
possible that for the recording of a power curve only a single data point is
registered, from
which ¨ if required ¨ a theoretical power curve can be approximated.
The invention has discovered that, in the case of the formation of ice on the
rotor blades of a
wind turbine, and the deterioration in the aerodynamics of the rotor blades
that often
accompanies it, even though the control of the blade angles has to be changed
to avoid
stalls, the changing of the blade angle control can be optimized iteratively
in order to
minimize as far as possible the loss of yield resulting from the formation of
ice. This applies
in particular because it has been found that, with the formation of ice on the
rotor blades,
completely different ice formations occur in each case, correspondingly
changing the
aerodynamics of the rotor blades individually. A basic static changing of the
blade angle
control, which ultimately is also what happens when a minimum blade angle is
prescribed, is
therefore disadvantageous with regard to the yield of a wind turbine. The
method according
.. to the invention is relevant in particular for those wind turbines on which
the permanent
formation of ice over a relatively long time period can be expected, or
turbines in areas with
low or negative temperatures over several months.
According to the invention, it is provided that, when the formation of ice is
found, a wind
turbine for which the blade angle setting is performed during normal operation
on the basis of
a standard characteristic curve, in dependence on a characteristic number that
can be
ascertained during the operation of the wind turbine, is operated at first on
the basis of a
4
CA 3009204 2018-06-22
prescribed initial special characteristic curve, but in the further course of
the method this
special characteristic curve is iteratively adapted in order in spite of the
formation of ice to
nevertheless achieve the maximum possible yield. The initial special
characteristic curve
generally deviates from the standard characteristic curve.
For detecting the formation of ice on the rotor blades, various methods are
known in the prior
art. For example, the detection of the formation of ice may take place by
finding there is a
power deficit of the wind turbine with respect to a reference power curve, in
dependence on
the wind speed found. The wind speed may in this case be found by means of an
anemometer. Preferably, the outside temperature is also taken into
consideration, such that
a formation of ice is only detected if the temperature is below 3 C. If the
power fed in by the
wind turbine drops in comparison with the reference power expected in view of
the wind
conditions, formation of ice on the rotor blades can be presumed, in
particular at
temperatures below 3 C. Alternatively or in addition, the formation of ice on
one of the rotor
blades can be presumed if it is found that an imbalance is occurring at the
temperatures
mentioned.
The special characteristic curve initially used for operating the wind turbine
in the case of a
detected formation of ice may be a fixed prescribed special characteristic
curve. However, it
is also possible that the initial special characteristic curve is a special
characteristic curve
that was optimized in the case of a previous formation of ice on the wind
turbines by the
method according to the invention explained still further below. If it can be
assumed that the
formation of ice on a particular wind turbine always takes place in a similar
way, the
optimization described below can possibly be carried out more quickly, i.e.
with fewer
iterations, by referring back to a special characteristic curve that was
optimized in a previous
case of the formation of ice.
Both the standard characteristic curve and the special characteristic curve
preferably
indicates the blade angle, in dependence on the tip speed ratio A of the wind
turbine. The
dimensionless tip speed ratio A is the ratio of the circumferential speed of
the rotor blades to
the wind speed and ¨ if not already calculated for other reasons ¨ can be
readily calculated
from the measured values with regard to the wind and the status of the wind
turbine that are
registered by the control of the wind turbines.
5
CA 3009204 2018-06-22
After the changeover of the operation of the wind turbine on the basis of an
initial special
characteristic curve for the blade angle, first a first power curve is
recorded for a prescribed
time period. The power curve in this case reflects the ratio of power fed in
to wind speed.
Given a sufficiently long time period in which different wind speeds can often
be expected,
sufficient data points can be obtained to ascertain the power curve
sufficiently accurately.
The length of the time period for recording the first power curve, but also
the power curves
subsequently to be recorded, may in principle be chosen as desired, and may
even be
several days, for example 3 days. It is preferred, however, if the length of
the time period in
question is preferably less than 24 hours, preferably less than 12 hours, more
preferably less
io .. than 6 hours. It is also possible that the recording of the first power
curve is only performed
for a time period of about 30 minutes, preferably of about 10 minutes.
Although the number
of ascertainable data points in a correspondingly short time period may not be
sufficient for
the complete determination of a power curve, it can generally be extrapolated
or
approximated by means of theoretical models if a complete power curve is
indeed required.
Subsequently, the special characteristic curve is changed. For this, it is
particularly preferred
if the special characteristic curve is parameterized by a parameter. A
changing of the special
characteristic curve can then be achieved simply by changing the parameter.
The parameter
may in this case be changed for example by a fixed prescribed increment. It is
however also
possible that the increment is changed in dependence on the characteristic
number that can
zo be ascertained during the operation of the wind turbine, in particular
therefore for example
the tip speed ratio A. Thus, the increment may for example be formed from the
characteristic
number that can be ascertained during the operation of the wind turbine
multiplied by a
constant factor.
After appropriate changing of the special characteristic curve, a further
power curve is
recorded for a prescribed time period. The recording is in this case performed
analogously to
the recording of the first power curve, for which reason reference is made to
the statements
given above.
It is subsequently checked whether the last ascertained further power curve
represents an
optimum. This check may be performed by using comparisons with the previously
ascertained power curves on the basis of the initial special characteristic
curve and/or
changed special characteristic curves.
6
CA 3009204 2018-06-22
In order that the comparison of the power curves is sufficiently meaningful,
the check as to
whether the last ascertained further power curve represents an optimum is
preferably only
carried out if the turbulences in the time period of the recording of this
power curve are
comparable to the turbulence in the time period of the recording of the
previous power curve.
If not, the recording of the further power curve is repeated. The turbulence
can in this case
be depicted for example by way of the turbulence intensity, that is to say the
ratio of the
standard deviation of the wind speed to the mean value of the wind speed.
Alternatively, it is possible to convert the ascertained power curves to
reference conditions
and perform the check as to whether the last ascertained further power curve
represents an
io optimum on the basis of these power curves converted for reference
conditions. The
conversion of the power curves to reference conditions provides a direct
comparability.
The check as to whether the last ascertained further power curve represents an
optimum
may be performed for example on the basis of yields that can be ascertained
from the power
curves and a prescribed wind distribution. For this purpose, the yield to be
expected during
operation with the corresponding power curve is calculated for a prescribed
wind distribution
and used as a criterion for comparison. If the power curves only comprise a
small number of
data points, or even only one data point each, the check may even be confined
to a direct
comparison of the electrical power generated ¨ in particular if the wind
speeds of the
individual data points of the individual power curves are comparable. In this
case, it is not
zo necessary to extrapolate or approximate a power curve for the data
points registered.
If it is found that the last ascertained further power curve is an optimum,
the wind turbine is
operated on the basis of the optimum special characteristic curve, which is
based on the last
ascertained further power curve. If an optimum has still not been reached yet,
the special
characteristic curve is changed once again and the method subsequently
repeated
iteratively.
In the case of the method described, the optimization of the special
characteristic curve may
be carried out for each individual blade one after the other, after
ascertaining the optimum
special characteristic curve for one rotor blade, the aforementioned steps a)
to e) being
repeated one after the other for the further rotor blades. If, for example, it
is presumed
because of an imbalance that ice has formed on a rotor blade and this rotor
blade can be
7
CA 3009204 2018-06-22
identified on the basis of the imbalance, it is preferred to carry out the
method for this rotor
blade first.
Alternatively, it is possible to set the blade angle of all the rotor blades
uniformly by means of
the initial and subsequently optimized special characteristic curve. In
particular in this case, it
is however preferred that, after an optimization of the common special
characteristic curve, a
blade angle correction is carried out for each rotor blade individually one
after the other.
For this purpose, first the special characteristic curve is changed for an
individual rotor blade.
The blade angle of this one rotor blade is therefore controlled differently
from the special
characteristic curve ascertained previously as common to all the rotor blades.
The changing
io of the special characteristic curve for an individual rotor blade may in
this case be performed
analogously to the previously described changing of the special characteristic
curve for all
the rotor blades, that is to say in particular by changing the parameter of a
parameterized
special characteristic curve. Alternatively, it is also possible that the
changing of the special
characteristic curve for an individual rotor blade is achieved by prescribing
a blade angle
deviation. The blade angle of the rotor blade is in this case set by a
constant Delta in
comparison with the special characteristic curve for all the rotor blades.
Subsequently, a power curve is again ascertained for a prescribed time period
and the
special characteristic curve for the individual rotor blade is changed
iteratively as long as it
takes until an optimum is achieved. The preconditions and/or possibilities for
checking for the
presence of an optimum for the special characteristic curve of an individual
rotor blade, what
was said with regard to the optimization of the special characteristic curve
for all the rotor
blades applies analogously.
If an optimum for the special characteristic curve of an individual rotor
blade is found, the
blade-individual optimization is carried out for the other rotor blades of the
wind turbine one
after the other.
After completion of the method described, a repetition of the method may be
envisaged. The
repetition may be initiated by ambient conditions that have perhaps changed
decisively (for
example temperature or atmospheric humidity) or for example take place on the
basis of a
prescribed time interval.
8
CA 3009204 2018-06-22
For an explanation of the wind turbine according to the invention and the
computer program
product according to the invention, reference is made to the statements given
above.
The invention is now described by way of example on the basis of an
advantageous
embodiment with reference to the accompanying drawings, in which
Figure 1 shows a wind turbine designed for carrying out the method
according to
the invention; and
Figure 2 shows a flow diagram of a first embodiment of the method
according to
the invention.
In Figure 1, a wind turbine 1, which is designed for carrying out the method
according to the
invention, is outlined.
The wind turbine 1 comprises a rotor 2 with a number of rotor blades 4 that
are adjustable in
blade angle by means of blade adjusting drives 3, which is arranged rotatably
on a nacelle 5.
The nacelle 5 is in turn arranged rotatably on a tower 6.
The rotor 2 drives via the rotor shaft a gear mechanism 7, which is connected
on its output
side to a generator 8. A wind-induced rotational movement of the rotor 2 can
thus be
converted into electrical energy, which can then be fed, possibly via
converters (not shown)
and/or transformers 9, into an electric grid 10.
The wind turbine 1 also comprises a control device 11, which is connected by
means of
control lines (not shown) to the various components of the wind turbine 1 in
order to control
them. As known from the prior art, the control device 11 also registers the
measured values
not only from sensors (not shown) of the wind turbine 1, which register
operational
characteristic numbers such as the rotor speed, but also from sensors 12 for
registering the
wind speed and temperature in the region of the nacelle. The control device 11
is also
designed to derive further characteristic numbers, such as the tip speed ratio
A, from the data
registered.
9
CA 3009204 2018-06-22
The control device 10 is designed inter alia to set the blade angle of the
rotor blades 4. In this
respect, the control device 11 is designed for running a computer program
product with
which the method explained below is carried out.
In normal operation, the setting of the blade angles of all the rotor blades 4
is performed on
the basis of a standard characteristic curve, by means of which the optimum
blade angle for
the operation of the wind turbine is ascertained on the basis of the tip speed
ratio A (step 90).
It is regularly checked on the basis of the tip speed ratio A, the measured
wind speed in the
region of the nacelle and the power actually fed in whether the momentary
yield deviates
from the theoretical yield of the wind turbine that can be ascertained by
means of a reference
io power curve (step 95). If this is the case and the temperature in the
region of the nacelle is
still below 3 C, the otherwise imminent further operation of the wind turbine
is interrupted on
the basis of the standard characteristic curve and a changeover is made to
step 100 of the
method. Yet other methods suitable for detecting a formation of ice on wind
turbines that can
be used equally well at this point are possibly known in the prior art.
However, the chosen
method has the advantage that the relies exclusively on measured variables
already
registered for other reasons concerning the control of a wind turbine, and to
this extent no
additional components are required.
In step 100, the operation of the wind turbine 1 is carried out of a special
characteristic curve
for the blade angles of the rotor blades 4 that is stored in a fixed state in
the control device
11. The special characteristic curve deviates from the standard characteristic
curve, but also
reproduces a relationship between the blade angle and the tip speed ratio A.
The special
characteristic curve is in this case parameterized, i.e. the profile of the
special characteristic
curve can be adapted by a parameter. For the initial special characteristic
curve, an initial
value for this parameter is established.
Subsequently, a first power curve of the wind turbine 1 is recorded in a known
way for a
prescribed time period of 12 hours (step 105). The recording of a
corresponding power curve
is known from the prior art.
Subsequently, the special characteristic curve is changed by changing its
parameter (step
110). The parameter is changed by an increment that is obtained from a product
of the tip
speed ratio A and a constant factor, the algebraic sign of the change being
obtained from
rules that are known for iterative methods.
CA 3009204 2018-06-22
Subsequently, a further power curve of the wind turbine 1 is once again
recorded in a known
way for a time period of 12 hours (step 115).
In the subsequent step 120, it is checked whether the turbulence intensity
during the
recording of the further power curve in step 115 is comparable to the
turbulence intensity
during a corresponding previous recording, for example the recording of the
first power curve
in step 105. This can be ascertained on the basis of wind data recorded by the
control unit
11. If the turbulence intensity is not comparable, once again a power curve of
the wind
turbine 1 is recorded on the basis of the changed special characteristic curve
(step 115). It
goes without saying that a completely new 12-hour time period does not have to
be begun
io for this. Rather, it is possible to continue the already previously
begun recording of the power
curve, and to limit the results in each case to the last 12 hours of the
recording, until the
turbulence intensity in these 12 hours specifically is comparable to the
turbulence intensity
during a corresponding previous recording.
If the turbulence intensity for the last recorded power curve corresponds to
the prescribed
values, it is checked whether the last recorded power curve represents an
optimum. For this
purpose, a theoretical yield is ascertained with the aid of a prescribed wind
distribution on the
basis of the last recorded power curve and is compared with the corresponding
yields of the
previously recorded power curves, including the first power curve. If the
yield of the last
recorded power curve is not optimum, a jump is made back to step 110, where
the special
zo characteristic curve is changed once again, whereupon the subsequent
steps 115 to 125 are
performed as described.
If the check in step 125 finds that the last ascertained power curve
represents an optimum,
the operation of the wind turbine is continued on the basis of the special
characteristic curve
taken as a basis for this power curve (step 130).
The optimized special characteristic curve thus ascertained is first used for
setting the blade
angles of all the rotor blades 4 of the wind turbine 1, whereby a yield that
is often higher in
comparison with the initial special characteristic curve can already be
achieved by the wind
turbine 1. In order to optimize this yield still further, an individual blade
angle correction is
also carried out individually for each rotor blade 4 on the basis of the
optimized special
characteristic curve found.
11
CA 3009204 2018-06-22
For this purpose, in step 135, the special characteristic curve for an
individual rotor blade is
changed by prescribing a constant blade angle deviation, which is applied to
the blade
angles for this one rotor blade actually ascertained by means of the special
characteristic
curve. It is however also possible to change the special characteristic curve
by means of the
adaptation of the respective parameters of the parameterizable special
characteristic curve
for this particular blade, in order in this way to optimize the energy yield.
Subsequently, a further power curve is recorded for a time period of 12 hours
(step 140). If
the turbulence intensity in this time period does not coincide with the
turbulence intensity of
the previously recorded power curve, the recording of the power curve is
repeated or
io continued until a power curve based on a time period of 12 hours with a
comparable
turbulence intensity could be found (step 145).
If there is a suitable power curve, it can be checked in step 150 in a way
comparable to step
125 whether this power curve represents an optimum. If this is not the case,
the special
characteristic curve for the rotor blade 4 in question is changed once again
according to step
135 and the subsequent steps are run through as described. If the power curve
ascertained
is the optimum, the control of the blade angle of the rotor blade in question
is carried out on
the basis of the ascertained individual special characteristic curve.
Subsequently, the method
is repeated as from step 135 for each rotor blade 4 of the wind turbine 1,
until there is an
individual special characteristic curve for each rotor blade 4.
The operation of the wind turbine 1 then takes place on the basis of the
special characteristic
curves optimized individually for each rotor blade 4.
In the case of the exemplary embodiment described above, the power curves are
recorded in
each case over a time period of 12 hours. It is however also possible to limit
the recording to
or 10 minutes in each case. Even if a complete power curve generally cannot be
25 ascertained in such a time period, the data point or points thereby
obtained may be sufficient
to carry out the optimization of the special characteristic curve.
12
CA 3009204 2018-06-22
